Cardiac pathologies are detected by one-dimensional electrocardiogram signals or two-dimensional images. When working with electrocardiogram signals, they can be represented in the time and frequency domains (one-dimensional signals). However, this technique can present hurdles, such as the high cost of private health services or the time it takes for the public health system to refer the patient to a cardiologist.

In addition, the variety of cardiac pathologies (more than 20 types) is a problem when it comes to diagnosing a condition. Moreover, surface electrocardiography (sECG) is a little explored technique for this diagnosis.

With this in mind, the team made up of Evelyn Aguiar, Biomedical Engineering graduate of the School of Biological Sciences and Engineering in collaboration with professors Diego Almeida, Ph.D., and Fernando Villalba, Ph.D, and student Daniel Amaguaña, all from the same school of School of Biological Sciences and Engineering; along with Andrés Tirado, Ph.D., professor of the School of Mathematical and Computational Sciences, developed the project Rapid Detection of Cardiac Pathologies by Neural Networks Using ECG Signals (1D ) and sECG Images (3D). The project is published in the scientific journal Multidisciplinary Digital Publishing Institute (MDPI), indexed in the Scopus database (Q2), focused on areas such as: computer science, mathematics, and modeling and simulation.

The main objective of the research was the detection of cardiac pathologies using LSTM and ResNet34 neural networks with surface electrocardiography (sECG) signals. sECGs are three-dimensional images (two dimensions in space and one in time). It is expected that the use of these models will serve as a tool to support the primary care physician in reaching a diagnosis of cardiac pathologies. This does not imply replacing health professional, but rather allowing a significant improvement in the diagnosis of cardiac conditions.